A Friendly Reminder that AToL Is Really ANoL C. Randal Linder University of Texas at Austin (currently the Program for Evolutionary Dynamics, Harvard University) Main Points • Reticulation (hybrid speciation and horizontal gene transfer/introgression) is so common in some groups that the AToL project should incorporate it from the outset. • Detection and reconstruction of reticulation is in its infancy. • Combined data analyses will require explicitly ruling out reticulation. • Systematics will need to become more intimate with population genetics. Unruly Nature Whatever is not forbidden will occur. -- Gerald Myers (ca 1980) In Other Words • If a set of genes can be brought together in a cell, survival and reproduction will be determined by the phenotype produced in the environment of the organism. • If the organism can survive and reproduce as well as or better than its competitors, it “works” no matter the mating/process that produced it Evidence of HGT in Bacteria • Relatively large proportions of bacterial genomes appear to be subject to horizontal gene transfer. • Recent evidence indicates there is probably a “core” set of genes that are unlikely to be involved in HGT (Daubin et al. 2003. Science 301:829). • Can these low transfer genes be routinely identified and used for reconstruction? – Need to look in order to know Hybrid Speciation in Eukaryotes • Primarily found in plants, but also some fish and amphibians – Also in some reptiles but these are usually sterile asexual lineages. – What’s up with fungi? • Most recent estimates suggest that about 7% of new fern lineages have arisen by hybrid speciation and 3-4% of angiosperm lineages (Otto and Whitten. 2000. Annu. Rev. Genet. 34:40) Current Methods for Detecting and Reconstructing Hybrid Speciation are in their Infancy Phylogenetic Trees: Models, Methods, and Results • Models of speciation: YES • Models of sequence evolution: YES • Tree distances and accuracy measures: YES • Methods and optimization problems for tree inference: YES Phylogenetic Networks: Models, Methods, and Results • Just two years ago: – Models of speciation: NO – Models of sequence evolution: NO – Network distances and accuracy measures: NO – Methods and optimization problems for network inference: YES/NO Where We Are Today • Models of speciation: yes • Models of sequence evolution: yes • Network distances and accuracy measures: yes • Methods and optimization problems for network inference: yes Reconstruction Quality Model Phylogeny: 20-taxon 1-hybrid network NeighborNet: False Positives Open Problems in Hybrid Speciation • Detecting reticulation • Representing reticulate evolutionary scenarios • Inferring reticulate evolution • Visualization Deciding Whether to Combine Datasets A B C D E A B C D E A B C D E Population Genetic Issues • Population genetic reticulation masquerading as hybrid speciation – Meiotic recombination – Sexual recombination • Lineage sorting • Resolving this will require combining coalescent models with phylogenetic models Lineage Sorting: Gene Tree/Species Tree Gene Tree/Species Tree Gene Tree/Species Tree Gene Tree/Species Tree Gene Tree/Species Tree.